Frequently Asked Questions
Common questions about internet speeds, what the numbers mean, and how to get a faster connection.
Why is my speed lower than I'm paying for?
Common causes include Wi-Fi (performance varies with the standard, channel width, signal, interference, and device), an old router, congestion at peak hours, a VPN, or your device's network card. Plans are also sold as "up to" speeds. Plug directly into your router with Ethernet for the clearest measurement of the connection reaching your home, then compare repeated tests with your plan's advertised rate.
Why do I get different results than Ookla Speedtest or fast.com?
Every test measures against different servers with different methods. Ookla may pick an ISP-hosted server nearby or inside your ISP's own network, while fast.com tests against Netflix's servers and starts with download only. We use a Cloudflare endpoint when its probe succeeds, but download and upload choose independently and either can fall back to our own backend. Differences are expected because paths, server load, and timing differ; a large gap is a reason to compare the same device, connection, and test conditions.
What internet speed do I actually need?
Less than most people think. One 4K stream needs 25 Mbps; a video call needs 3–5 Mbps down and up; online gaming needs under 3 Mbps but low, stable ping. A household of four streaming, gaming, and working simultaneously is comfortable at 100–300 Mbps. Beyond ~300 Mbps, extra download speed rarely changes everyday experience — latency and upload matter more. Use our free calculator to get a number for your household.
Why is my upload so much slower than my download?
Most cable and DSL plans are asymmetric by design — upload is typically 5–35 Mbps even on a 1 Gbps download plan, because the cable spectrum is allocated mostly to downstream. Only fiber (Verizon Fios, AT&T Fiber, Google Fiber) and the top tier of some newer cable plans deliver symmetric speeds. Slow upload is what makes video calls stutter and cloud backups crawl.
What's a good ping?
This test's grade bands treat HTTP latency to the selected endpoint of 10 ms or less as A+, 25 ms or less as A, and 60 ms or less as B. Those timings include browser, endpoint, route, and connection effects. A page, game, or call may use another endpoint and route, so this observation does not predict its performance.
What is jitter?
Jitter is the variation in ping over time. Low jitter (under 5 ms) means a stable connection — useful for video calls and gaming. High jitter (over 30 ms) can coincide with dropped frames or rubber-banding, even if your average ping looks fine. Jitter can have several causes, including Wi-Fi conditions, congestion, the route, or the test endpoint; a result alone cannot identify an ISP or equipment fault.
What is packet loss?
Packet loss is the percentage of network packets that do not reach their destination. This browser HTTP test does not directly measure IP packet loss, so current results show it as not measured. Use a suitable packet-level tool when you need that measurement; HTTP request failures are not a substitute.
What is loaded HTTP latency, and why did it lower my test grade?
Loaded HTTP latency is the observed increase in round-trip time to the selected download endpoint while download or upload transfers run. The displayed test grade applies published bands to that observation. Browser, endpoint, route, Wi-Fi, and queueing effects can all contribute, so the result does not diagnose bufferbloat or predict how a call or game will perform. Compare repeated Ethernet and Wi-Fi tests before changing equipment or plans.
Why did I get a bad grade with a fast connection?
The observed test grade is the worst available sub-grade among download, upload, HTTP latency, HTTP jitter, and loaded HTTP latency; an unavailable metric is omitted. It describes this run against our published bands, not your internet service overall and not future application performance. The result card names the observation that set the grade and suggests checks to investigate.
Should I test over Wi-Fi or Ethernet?
Both, for different questions. Ethernet provides a useful comparison for the connection reaching your home; Wi-Fi measures what your devices actually experience. If Ethernet is fast but Wi-Fi is slow, investigate the router, its placement, interference, and the device, but do not treat that result alone as proof of a provider problem.
How can I make my internet faster without upgrading my plan?
The biggest wins are often free: move your router out of a cabinet and up off the floor, switch congested devices to 5 GHz or Ethernet, reboot a modem that has been up for months, check a router older than ~5 years for firmware and compatibility before replacing it, and turn off a VPN when you don't need it. In multi-story or large homes, a mesh system may help with dead zones when coverage is the limiting factor; test placement and wiring before buying equipment.
Is my ISP throttling me?
Maybe — but congestion and Wi-Fi problems are far more common than deliberate throttling. Real throttling shows a pattern: speeds fine over a VPN but slow without one, or specific services (video, torrents) slow while speed tests look normal. Run tests at different times of day, over Ethernet, with and without a VPN, and compare. Consistent slowness at 7–11 pm is neighborhood congestion, not throttling.
How accurate is a browser-based speed test?
A browser-based result depends on the browser, device, parallel HTTP requests, route, and server conditions, so it is not a guarantee of an exact line rate—particularly on very fast connections. A wired iperf3 test can provide another reference when you need one. We use a Cloudflare endpoint when its probe succeeds, but each direction can independently fall back to our own backend, and the selected endpoint is not guaranteed to be the nearest.
How much data does the test use?
Data use is not a fixed amount. Download and upload each run a 2-second warmup followed by an 8-second measurement window with parallel workers, so usage depends on the speed reached and can exceed 1 GB on fast connections. Take care on a metered plan.
Is my data private when I run a test?
A test does not require an account, and its measurements are not saved to our results database unless you create a shareable link. Our server and location providers may process your IP address, and hosting services may process request data. Google's privacy scripts can load before you make a choice; manual article ads are requested only after Google supplies consent information. See the Privacy page for details.
Measurement details verified against implementation on .